| Summary: | Tools are needed that can add value to existing drought information and customize it for specific drought management contexts. This study develops a generalized framework that can be used to link local impacts with readily available drought information, thus increasing the usability of existing drought products in decision making. We offer a three-step risk-based framework that can be applied to specific decision-making contexts: (i) identify hydrologic impact thresholds, (ii) develop threshold exceedance model, and (iii) evaluate exceedance likelihood. The framework is demonstrated using a study site in south-central Oklahoma, which is highly susceptible to drought and faces management challenges. Stakeholder input from interviews are used to identify “moderate” and “extreme” thresholds below which water needs are not met for important uses. A logistic regression model translates existing drought information to the likelihood of exceeding the identified thresholds. The logistic model offers an improvement over climatology, and the 12-month Standardized Precipitation Index is shown to be the best drought index predictor. The logistic model is used in conjunction with historical drought information to give a retrospective look at the risk of drought impacts from the beginning of the century. Results show the 1980s to early 2000s to be an anomalously wet period, and that recent drying trends and impacts do not appear to be unusual for the 20th century. This drought risk analysis can be used as a baseline by local managers to guide future decision making under climate uncertainty.
|
|---|
